Hetzner charges EUR 4.55/month for a 2 vCPU, 4GB RAM VPS. DigitalOcean's equivalent runs $12/month. AWS on a Savings Plan runs $12-24/month for similar specs. That price gap is not a rounding error. For teams running a dozen self-hosted services, it's the difference between a EUR 50/month infrastructure bill and a $300/month one.
The tradeoff is specific: Hetzner has a thinner managed-service catalog, slower support response times, a primarily European footprint, and a smaller ecosystem of third-party integrations. Those limitations rule it out for certain workloads. For everyone else, the pricing is hard to argue with.
Current Pricing
| Instance | Provider | Cost/Month | Specs |
|---|---|---|---|
| CX22 | Hetzner | EUR 4.55 | 2 vCPU AMD, 4GB RAM, 40GB NVMe |
| CAX11 (ARM) | Hetzner | EUR 3.79 | 2 vCPU ARM, 4GB RAM, 40GB NVMe |
| AX41 Dedicated | Hetzner | EUR 39.00 | 6-core/12-thread, 64GB RAM, 2x512GB |
| Basic Droplet | DigitalOcean | $12.00 | 2 vCPU, 2GB RAM, 50GB SSD |
| Premium Droplet | DigitalOcean | $18.00 | 2 vCPU, 4GB RAM, 80GB SSD |
| t4g.small | AWS (Savings Plan) | ~$12.00 | 2 vCPU, 2GB RAM |
| t4g.medium | AWS (Savings Plan) | ~$24.00 | 2 vCPU, 4GB RAM |
Storage compounds the gap. Hetzner block storage costs EUR 0.0476/GB/month. AWS EBS gp3 runs $0.08/GB/month in US regions. Hetzner S3-compatible object storage is EUR 0.0119/GB/month; AWS S3 standard is $0.023/GB/month.
Egress is where AWS billing really separates from Hetzner. AWS charges $0.09/GB for outbound traffic after the free tier. Hetzner includes generous monthly traffic quotas with each instance and does not charge per-GB egress within quota. A VPS serving 10TB/month of outbound data pays nothing extra on Hetzner. On AWS, that's approximately $900 in transfer charges on top of the instance cost.
What Hetzner Gets Right
ARM instances perform well. The CAX series uses Ampere Altra processors. For CPU-bound workloads like compiling code, running containers, or serving web traffic, the CAX11 at EUR 3.79/month benchmarks above most $12/month VPS options on single-core tasks. Teams running CI/CD pipelines across fleets of build runners save significantly by moving those runners to Hetzner ARM instances.
Dedicated servers are priced practically. Hetzner's AX41 at EUR 39/month gives 64GB RAM and dual NVMe storage. AWS bare metal instances start at several dollars per hour. For database servers, ML inference workloads, or rendering needing predictable performance without noisy neighbors, Hetzner's dedicated tier has no real equivalent at a similar price point.
EU data residency comes standard. Hetzner operates in Germany, Finland, and (since 2023) Ashburn, Virginia. European companies with GDPR data residency requirements get compliant EU hosting at standard pricing. Cloud providers that market "EU-compliant" tiers typically charge 10-20% more for the same hardware.
Billing is predictable. No surprise egress charges, no per-request fees on load balancers, no data transfer costs between instances in the same datacenter. Finance teams dealing with variable AWS bills appreciate the simplicity.
Object Storage and Backup Workloads
Object storage pricing is one of the clearest cases for Hetzner over AWS. Hetzner's S3-compatible object storage at EUR 0.0119/GB/month is roughly half the cost of AWS S3 standard at $0.023/GB/month. A team storing 20TB of backup data pays roughly EUR 238/month on Hetzner versus approximately $460/month on AWS S3, before considering AWS data retrieval costs.
Hetzner object storage is compatible with the AWS S3 API, meaning most tools that work with S3 (rclone, restic, Duplicati) work with Hetzner's service without code changes. The switch is often a single endpoint URL configuration change. For teams storing database backups, static assets, user-uploaded files, or media archives, this cost difference compounds quickly.
What Hetzner Gets Wrong
Support is slower. Hours rather than minutes is realistic for non-critical tickets. There is no enterprise support tier with contractual SLAs. When production fails at 2 AM, Hetzner's support model creates friction that AWS and DigitalOcean both handle better.
The managed services catalog is thin. No managed Kubernetes with the operational maturity of EKS or GKE. No managed relational databases with automatic failover comparable to AWS RDS. No CDN, serverless compute, or ML infrastructure. Hetzner provides raw compute and storage. Everything on top of that is your team's responsibility.
US geographic reach is limited to a single Ashburn region. Teams requiring multi-region North American redundancy need a different provider.
Third-party ecosystem depth is smaller. AWS has mature Terraform providers, a large partner network, and integrations with nearly every developer tool. Hetzner's ecosystem is growing but thinner. Teams migrating existing AWS Terraform configurations should budget engineering time for the transition.
DigitalOcean's Current Position
DigitalOcean's original pitch was "AWS without the complexity." That positioning has weakened. AWS simplified its interface considerably over the past three years, and DigitalOcean raised prices to the point where the gap versus AWS for equivalent specs is now modest. Their App Platform and Managed Kubernetes are solid products, and the developer experience remains smoother than AWS for common use cases. But Hetzner undercuts them on raw compute cost, and AWS outperforms them on managed service depth.
DigitalOcean fits teams wanting managed Kubernetes or App Platform without full AWS operational investment, and who value a smoother developer experience over the lowest possible infrastructure cost.
When AWS Is Still Worth It
Managed services with enterprise SLAs: RDS Multi-AZ, EKS managed control plane, CloudFront, Aurora Serverless. None have real equivalents at Hetzner's price point. You're paying for operational complexity handled by someone else.
US enterprise and government compliance: FedRAMP authorization, HIPAA Business Associate Agreements, and PCI-DSS Level 1 from AWS satisfy requirements that Hetzner's certifications do not cover for US buyers.
Global multi-region architecture: AWS operates 30+ regions. Hetzner has 5 locations. Active-active redundancy for globally distributed users requires AWS's geographic coverage.
Existing team investment: If your engineers are AWS-certified and your infrastructure is AWS Terraform, retraining and rebuilding often costs more than infrastructure savings justify.
For a full comparison of the hyperscaler tier, see our AWS vs Google Cloud vs Azure breakdown.
Infrastructure as Code and Tooling
Hetzner has a well-maintained Terraform provider that handles VMs, floating IPs, load balancers, firewalls, and volumes. Teams already using Terraform can manage Hetzner infrastructure with the same tools they use for AWS or GCP. The provider is actively maintained and covers the full Hetzner Cloud API.
Migrating existing AWS Terraform configurations to Hetzner is not a drop-in replacement. AWS resource types all have Hetzner equivalents, but the resource definitions need to be rewritten. For a complex infrastructure with dozens of resources, budget a week of engineering time for migration. For a simple staging environment with a handful of VMs, a day is realistic.
Ansible, Chef, and Puppet all work on Hetzner instances identically to any other Linux VPS. GitHub Actions and GitLab CI runners can target Hetzner VMs directly. Docker and Kubernetes run without modification. The tooling ecosystem for operating workloads on Hetzner is not a constraint. The constraint is managed services: anything AWS was running for you, you now run yourself.
One specific callout: Hetzner load balancers work well for straightforward HTTP and TCP load balancing. They do not offer the depth of AWS Application Load Balancer routing rules, header-based routing, or WAF integration. Teams relying on ALB advanced routing features need to evaluate whether nginx, HAProxy, or Traefik covers their use case before migrating.
The Hetzner Cloud Console is clean and functional, though it has fewer configuration options visible at a glance than AWS. For teams that find the AWS Console overwhelming, Hetzner's simplicity is a feature. For teams that need complex networking configurations or fine-grained IAM policies, the simpler interface sometimes becomes a limitation.
The Recommendation
Hetzner is the right choice for personal projects, self-hosted services, EU data residency requirements, cost-constrained startups, and workloads that don't need managed databases with automatic failover or global CDN coverage.
DigitalOcean fits teams wanting managed Kubernetes and a smoother developer experience without AWS's operational overhead.
AWS fits production workloads with enterprise compliance requirements, teams already invested in the ecosystem, and anyone needing managed services that absorb operational complexity. The premium pays for real things. Make sure your workload needs those things before paying for them.